Questions about real prompting decisions
Every question is a choice you make when you write a prompt: which instruction to give, where a prompt goes wrong, what to change to fix it.
Features
TokIQ is deliberately simple: one question at a time, an explanation for every answer, and progress you can see. Here is everything in the app.
Every question is a choice you make when you write a prompt: which instruction to give, where a prompt goes wrong, what to change to fix it.
After you answer, tap Why? for a short explanation of the idea behind the question, followed by three follow-up questions on the same point.
Fundamentals, role & context, examples, reasoning, structured output, grounding, tools & agents, and safety.
Beginner, intermediate and experienced levels, so the quiz stays useful whether you are new to AI or ship LLM features for a living.
Tell TokIQ which tools you use (ChatGPT, Claude, Gemini, Copilot, Perplexity) and what for: work, study, writing, research or code.
See your accuracy topic by topic, so you know exactly which skills are still weak.
Compare your XP with other players on a global leaderboard, shown under a nickname you choose or a masked label.
An optional reminder in the evening, so a couple of questions a day becomes a habit.
No timer and no fixed round length. Answer one question or twenty, and set a daily goal of 3, 5, 10 or 20 if you want one.
The interface, questions and explanations are available in 40 languages, from English and Spanish to Azerbaijani, Japanese and Arabic.
A calm, low-glare design that is easy to read for longer sessions.
Open the app and start answering. Create an account with email, Google or Apple when you want to keep your progress.
Pick a topic
Short quizzes on how to write prompts that work.
Your app parses the model’s JSON, but sometimes the reply starts with “Sure! Here is the JSON:”. What is the most robust fix?
That’s right! +10 XP
Structured-output modes constrain the reply to valid JSON. Validation catches the rest.
Why string hacks break JSON parsing
A model that adds a friendly sentence before its JSON is doing what chat models are trained to do. Telling it “no preamble” helps, but only a schema-constrained mode makes the shape a guarantee.
Validate every reply. When it fails, retry once and include the validation error.
3 follow-up questions
Progress
68%
Accuracy
6
Day streak
1,240
XP
By topic
Plans
The free plan gives you new questions every day. Premium opens every topic and unlimited questions, gives you many more explanations and removes ads. Prices are shown in your local currency in the App Store and Google Play before you subscribe.
Free
Everything you need to start
Premium
All of TokIQ
Short quizzes on real prompting decisions, with an explanation for every answer. Free to start on iPhone and Android.